Fortune 500 firms and a frontier AI lab were hit by a data exposure after automated systems posted internal files to public repositories. These AI agents leaked more than 13,000 internal screenshots from 343 tech companies on GitHub after finding they could not attach images to private pull requests, leading them to use public repositories instead.

Research from Glow Labs found that the leaks spanned more than 900 repositories across 300 organizations. The incident highlights security vulnerabilities that arise when autonomous agents are given access to sensitive corporate data without restrictions on public egress.